Spatial binning EDA
Data used:
## [1] "../data/cleaned/prelim_singlereview.csv"Summary
From the preliminary data:
- Lots of studies with ‘unk’ spatial scale (this means that no plot nor site scales were reported by coders).
- Lots of small scale studies, some large scale studies
- Relatively few medium scale, and medium-large scale papers
- Not that many papers cross the scale bins we’ve used
My takeaway (Grant): A fair amount of small scale studies reach up into the medium scale category, but hardly any large scale studies have smaller scale components. Way too many studies do not report (or we couldn’t find) the spatial scale at which they are operating…maybe a lot of these are simulations? - doesn’t look like it, see below
Spatial binning to fit scale figure
The end result of these numbers may just be displayed as n=X in the existing scale figure, but there are other potential ways to display these things, too.
Key for binning:
## Scale_binned scales
## 1 Small 25m,50m,100m,500m,1km
## 2 Medium 10km,100km,1000km
## 3 Large 100Mgm,100Gm,101GmNumber of papers in each bin:
## # A tibble: 8 x 2
## Scale_groups count
## <fct> <int>
## 1 Small 86
## 2 Small,Medium 23
## 3 Medium 18
## 4 Medium,Large 6
## 5 Large 25
## 6 Small,Large 5
## 7 Small,Medium,Large 1
## 8 unk 71
Unknown scale studies…
Unclear what explains why so many studies had unknown spatial scale…
Small scale studies…
And large scale studies
Just for those studies with a ‘Yes’ for Q8: Does the paper consider or compare multiple spatial scales?
## Scale_binned scales
## 1 Small 25m,50m,100m,500m,1km
## 2 Medium 10km,100km,1000km
## 3 Large 100Mgm,100Gm,101GmNumber of papers in each bin:
## # A tibble: 8 x 2
## Scale_groups count
## <fct> <int>
## 1 Small 15
## 2 Small,Medium 3
## 3 Medium 4
## 4 Medium,Large 2
## 5 Large 10
## 6 Small,Large 2
## 7 Small,Medium,Large 1
## 8 unk 9